Complicated Diverticular Disease

Abstract

The modern management of colonic diverticular disease involves grouping patients into uncomplicated or complicated diverticulitis, after which the correct treatment paradigm is instituted. Recent controversies suggest overlap in management strategies between these two groups. While most reports still support surgical intervention for the treatment of complicated diverticular disease, more data are forthcoming suggesting complicated diverticulitis does not merit surgical resection in all scenarios. Given the significant risk for complication in surgery for diverticulitis, careful attention should be paid to patient and procedure selection. Here, we define complicated diverticulitis, discuss options for surgical intervention, and explain strategies for avoiding operative pitfalls that result in early and late postoperative complications.

Diverticulitis is broadly divided into uncomplicated and complicated diseases, the latter including complications of free perforation producing purulent or feculent peritonitis, abscess, fistula, obstruction from stricture, or chronically recurrent symptoms. This division is important as retrospective studies show that two or more episodes of uncomplicated diverticulitis are not associated with worsened outcomes. ¹ However, the likelihood of urgent operation increases by a factor of two with each subsequent hospitalization for complicated diverticulitis. In addition, morbidly obese patients and those younger than 50 years of age with complex diverticulitis are more likely to suffer a recurrence requiring intervention. ² When discussing patient options, it is important to allay concerns about acute worsening of disease that doesn't involve uncontained perforation. Conversion to perforated diverticulitis requiring emergent surgery is less than 6%, ^{3 4} with need for emergent stoma formation at 0.05%. ⁵ While an increasing rate of attacks can foreshadow fistula or abscess formation, this process is not associated with higher rates of emergent surgery or stoma formation. ⁵ Despite these reassuring data, chronically immunosuppressed patients are unique and should be considered for earlier elective surgery. ⁵ As surgical intervention is more common with complicated diverticulitis, and subsequent procedures are associated with worsened postoperative outcomes, careful patient-centered decision making is advocated.

Etiology and Presentation of Complicated Diverticulitis

Uncontained Perforated Diverticulitis

Approximately 22% of patients experience recurrence following index diverticulitis event and 55% experience recurrent symptoms following a second episode. ⁶ The risk of lifetime recurrence is higher in younger individuals and women; however, recurrence is not associated with higher risk of perforation. ⁶ Smokers demonstrate an increased risk of perforation and larger diverticular abscess formation compared with nonsmokers. ⁷ Nonsteroidal anti-inflammatory drugs and steroids both increase the risk of perforated diverticulitis. ^{8 9} In contrast, statins are associated with a decreased risk of diverticular perforation. ¹⁰ Knowledge of patient history can help determine risk of perforation.

Severity of diverticulitis is widely assessed and categorized using the modified Hinchey classification. ^{11 12} Uncomplicated diverticulitis includes stage 0 (clinically mild diverticulitis) and stage Ia (pericolic inflammation), which are successfully treated nonsurgically in 70 to 100% of cases. ^{13 14} Complicated diverticulitis includes stage Ib (abscess <5 cm in the proximity of primary inflammation), stage II (intra-abdominal, pelvic or retroperitoneal abscess, or abscess distant from the primary inflammation), stage III (generalized purulent peritonitis), and stage IV (fecal peritonitis).

The clinician's goal at initial assessment for suspected perforated diverticulitis is determining patient stability and need for emergent intervention. History usually reveals an abrupt onset of left lower quadrant or diffuse abdominal pain, often sharp and gnawing in nature and worsened by PO (Per Os) intake. Patients commonly report fevers, chills, anorexia, and fatigue. Some patients may describe a change in bowel habits involving more liquid or loose-quality stools. Blood is not commonly seen per rectum in patients with acute diverticulitis. Although only 1 to 2%of all patients with acute diverticulitis present as the Hinchey III or IV; postoperative mortality rates approach 20%, necessitating quick diagnosis and surgical management. ^{15 16}

Laboratory tests, such as white blood cell (WBC) count and C-reactive protein (CRP), may assist in monitoring patient response to therapy but normal values do not rule out the perforated diverticulitis. Plain abdominal films can quickly evaluate for pneumoperitoneum, while a computed tomography (CT) scan in stable patients provides additional information to include severity of pericolic inflammation, abscess location and size, adjacent organ involvement, and ruling out alternative diagnoses such as perforated ulcer, appendicitis, or tuboovarian abscess. CT scan is superior to contrast enema in its sensitivity (98 vs. 92%) with demonstration of bowel wall thickening and fat stranding, confirming active inflammation. ^{17 18} Barium enema is avoided in acute diverticulitis due to risk of peritoneal contamination.

Contained Perforated Diverticulitis

In contrast to uncomplicated diverticulitis managed in an outpatient setting, ¹⁹ patients with the Hinchey stage Ib or II are classified as contained perforated diverticulitis and often require hospitalization due to variable systemic toxicity, high fever, severe localized abdominal tenderness, and leukocytosis. Treatment of small mesocolic abscesses (<4 cm) with broad-spectrum antibiotics is successful in up to 70% of patients. ²⁰ Anaerobes (including Bacteroides , Peptostreptococcus , Clostridium , and Fusobacterium species ) are the most commonly isolated organisms, with Escherichia coli the most common gram-negative isolate. ^{21 22} Nationally, the most prescribed treatment is ciprofloxacin plus metronidazole for 10 to 14 days depending on patient response. Clinical improvement is often observed within 3 to 4 days of treatment initiation with transition to PO antibiotics when tolerating a diet.

Compared with small mesocolic abscesses (< 4cm), larger abscesses require percutaneous drainage and semielective surgery if drainage is unsuccessful. Large diverticular abscesses are associated with a high risk of recurrence, with successful drain placement often serving as a bridge to surgical resection. ²⁰ In a recent study of 511 complicated diverticulitis patients, 42% of them undergoing percutaneous drainage had recurrence of abscess and an increased probability of surgical intervention. ²³ As such, surgery is considered as rescue therapy for patients who deteriorate, fail to clinically improve, or have a persistent intra-abdominal abscess following drainage. ²⁴

Diverticular-Associated Fistula

A colovesical fistula is an abnormal connection between the colon and urinary bladder ( Fig. 1 ). In patients with history of diverticular disease, 2 to 18% demonstrated colovesicular fistula ²⁵ and sigmoid diverticulitis accounted for 65 to 90% of all fistulas. ^{18 26} Fistula rates are higher in males than females as the uterus, when present, provides protection from sigmoid inflammation. ²⁷ Symptoms of colovesical fistula include pneumaturia, hematuria, urinary frequency, and fecaluria. Patients often present with recurrent polymicrobial bladder infections and rarely, pyelonephritis. Cross-sectional imaging, cystoscopy, and colonoscopy are utilized for the evaluation of suspected colovesical fistula. ¹⁷ CT scan is the study of choice with presence of air in the noninstrumented bladder pathognomonic for fistula. ¹⁷ Colonoscopy is discouraged for evaluation less than 6 weeks from initial diverticulitis diagnosis, but is absolutely indicated upon resolution to rule out malignancy or inflammatory bowel disease. ²⁸ Colonoscopy is a poor test for diagnosis of a diverticular-associated fistula. While cystoscopy often reveals focal mucosal inflammation near a fistula and can rule out malignancy, visualization of the opening is uncommon. Contrast cystography can aid in visualization of a colovesicular fistula when other methods fail.

Fig. 1.

Colovesical fistula. CT scan showing diverticulum abutting a bladder with air. CT. computed tomography.

Colovaginal fistulas are abnormal epithelialized connections between the bowel lumen and vagina ( Fig. 2 ), accounting for 25% of all diverticular fistulas and most often seen in women who have undergone hysterectomies where the diseased colon is adherent to the vaginal cuff. ²⁹ Patients present with foul vaginal discharge or stool emanating from the vagina. True rectovaginal fistula is less common from diverticular disease and more typically results from inflammatory bowel disease, malignancy, trauma, or iatrogenic injury. Evaluation includes cross-sectional imaging or vaginogram.

Fig. 2.

Colovaginal fistula. Study showing contrast from vagina traveling into adjacent sigmoid colon.

Colocutaneous fistula involves communication between the bowel lumen and the skin ( Fig. 3 ), often the result of percutaneous abscess drainage. When not related to drain placement and occurring after surgical resection, they are typically attributed to an unrecognized anastomotic leak. ³⁰ Presence of a colocutaneous fistula, especially in and around the left flank and hip, in a patient without prior colonic resection should prompt evaluation for diverticulitis or colonic malignancy as a source.

Fig. 3.

Colocutaneous fistula. CT scan showing diverticulitis with inflammatory tract open to skin. CT. computed tomography.

Diverticular Stricture and Obstruction

Malignant diverticulitis describes a form of persistent, extensive inflammation, and a tendency toward obstruction and fistula formation. ²² Edema can result in a functional obstruction that usually resolves with antibiotic therapy and bowel rest. A chronic stricture occurs following recurrent attacks of diverticulitis due to pericolic fibrosis. ³¹ Patients presenting with colonic stricture often have a history of diverticulitis, complaints of obstructive symptoms, and are frequently older females. ²⁹ Workup with colonoscopy risks perforation and cannot always exclude malignancy, necessitating resection to obtain a diagnosis and relieve symptoms. While colonic stents can be used to relieve acute obstruction, they are considered a bridge to definitive surgical resection.

Operative Management of Complicated Diverticular Disease

Emergent or Urgent Management of Diverticulitis

The DILALA randomized trial comparing laparoscopic lavage (LL) to open Hartmann's procedure (OHP) for perforated diverticulitis demonstrated an increased 30-day mortality rate in the LL group (7.7 vs. 0%) and similar 90-day mortality (7.7 vs. 11.4%). ³² Despite no differences in number (52% lavage vs. 40% OHP) or severity of complications between groups, higher readmission was reported in the OHP group (0 vs. 5.7%). With no difference in reoperation rates, LL was deemed noninferior for control of the Hinchey III process. The SCANDIV trial evaluated LL versus laparoscopic or open resection with or without primary colon anastomosis. ³³ A 90-day postoperative complication (30.7% LL vs. 26% resection) and mortality rates (13.9% LL vs. 11.5% resection) were similar. Reoperation was significantly higher in the LL group (20.3%) compared with the resection group (5.7%), although reasons for reoperation were not provided. Additionally, four sigmoid cancers were missed on LL patients, supporting the case for colonoscopy in all nonresected patients. In a meta-analysis of the DILALA, SCANDIV, and LADIES trials, the main conclusion was LL produced a significantly lower 12-month reoperation rate (odds ratio [OR] = 0.32), at a cost of increased 90-day morbidity (OR = 1.7) attributable mostly to higher rate of intra-abdominal abscess formation (OR = 3.5). ³⁴ While LL failed in 17% of patients, no differences in mortality or morbidity were noted in those requiring subsequent surgical resection. LL by experienced surgeons may be considered in the Hinchey III diverticulitis, possibly averting a colostomy ^{24 35 36 37 38 39} at the cost of increased persistent/recurrent abdominal sepsis risk. ⁴⁰

Historically, the open creation of an end colostomy or HP was most commonly performed in the Hinchey III or IV patients, producing high rates of morbidity (25–75%) and mortality (2–30%). Additionally, fewer than 50% of patients underwent successful stoma reversal. ⁴¹ As a response to these data, resection with primary anastomosis was advocated in the setting of the Hinchey III or IV diverticulitis. Two randomized clinical trials have shown resection and primary anastomosis noninferior to HP. ^{41 42} Whether performing emergent HP or primary anastomosis, a laparoscopic approach showed no differences in 30-day morbidity or mortality compared with open HP. ⁴³ While effective in the right hands, remember that “in all cases, the adoption of LL in emergent settings, abdominal exploration for generalized peritonitis, pelvic dissection in inflammatory conditions, and possible suture of a diseased colon require the surgeon to have a minimum of colorectal and minimally invasive skills.” ⁴⁴

In a further attempt to limit high-morbidity operations for perforated diverticulitis, small case series have shown successful observation of free-intraperitoneal air in the clinically stable patient. ⁴⁵ Advocates of this nonoperative approach to free air argue treatment of the patient over the radiology; however, patients who have undergone organ transplantation, have HIV/AIDS, are taking corticosteroids, or are otherwise immunosuppressed are less likely to present with classic signs and symptoms of peritonitis, more often fail nonoperative management, decompensate rapidly, and have higher postoperative complications and death following delayed intervention. ^{10 46 47}

Elective Resection for Diverticulitis

The SIGMA trial, a multicenter, double blind, parallel-arm study of 104 patients between 2002 and 2006 demonstrated laparoscopic colectomy for diverticular disease produced fewer postoperative complications compared with open colectomy (15.4% absolute reduction) and no difference in mortality. ⁴⁸ A 10% leak rate in open cases, 6% rate in laparoscopic cases, and a 19.2% conversion rate (9.6% hand assisted and 9.6% open) confirmed the high-risk nature of colectomy for diverticulitis. Laparoscopy was associated with less pain and higher quality of life scores in the short term, while 6-month follow-up showed no difference in late complications or recurrence. ⁴⁹ In a financial analysis, the cost of a 1% complication reduction attributable to laparoscopy was a mere €31, making laparoscopic approach a bargain for patients and hospitals. ⁴⁹ A subsequent study in 2010 with similar design showed no difference in complication rates between laparoscopic (13.5%) and open (9%) groups, reduced perception of pain in laparoscopy and similar quality-of-life scores aside from cosmesis. ^{50 51} Patients undergoing conversion experienced worsened outcomes, making patient selection an important factor for better outcomes. ⁵²

Techniques for Improving Surgical Outcomes in Complicated Diverticulitis

Ureteral Stents

Most ureteral injuries occur at the base of the inferior mesenteric artery (IMA) and over the iliac vessels during dissection into the pelvis. American Society of Colona and Rectal Surgeons (ASCRS) practice parameters state ureteral stent use is at surgeon discretion and not favored due to low risk of injury (0.24–1.95%) in colorectal cases. ^{5 53} Recent data from National Surgical Quality Improvement Program (NSQIP) evaluating almost 60,000 colectomies from 2012 to 2014, demonstrated prophylactic ureteral stent placement (PUSP), were associated with a lower ureteral injury (UI) rate (OR = 0.59; 95% confidence interval [CI]: 0.41–0.84). ⁵⁴ Because stent placement prolongs operative time and increases cost, ⁵⁴ many providers avoid stent usage even in complicated cases such as diverticulitis and repeat pelvic surgery. A recent analysis of the National Inpatient Sample (NIS) indicated injury rates increased from 2.5 in 1,000 to 3.1 in 1,000 between 2001 and 2010, with higher rates in teaching hospitals (3.4 in 1,000), urban hospitals (2.8 in 1,000), and elective cases compared with emergent (3.0 vs. 2.4 in 1,000). ⁵³ Interestingly, rate of UI during colectomy for diverticulitis was 2.9 versus rectal cancer at 7.1 in 1,000. One hypothesis for the rising rate of injury is the increased use of laparoscopy ⁵⁵ ; however, studies are contradictory on this as a cause. ^{53 56} Supporting a laparoscopic approach, open procedures were recently associated with higher injury rates (OR = 1.31; 95% CI: 1.03–1.68). ⁵⁴ Risk stratification using imaging, pathology, and patient complaints is necessary to determine benefit from PUSP. Aside from prevention of UI, multiple studies suggest earlier identification of UI with stent placement. With the availability of newer technology, such as lighted stents, some strongly advocate for PUSP to improve both injury prevention and recognition. ⁵⁷

Ureteral stent placement results in minimal patient risk. Hematuria occurs in almost 100% of patients, especially if infrared (IR)/lighted catheters are utilized. ^{57 58 59} Given the high rate of hematuria with IR catheters, some authors express concern about thermal injury to ureters with risk of subsequent scarring, ⁵⁷ although, to date, there are no such injuries reported. Postoperative urinary tract infection (UTI) is linked to stent placement, occurring 0 to 5% of the time ^{57 58} ; however, PUSP was not independently associated with higher rates of UTI in an NSQIP cohort undergoing colectomy. ⁵⁴ Rate of urinary retention postoperatively is less than 5%. ⁵⁷ When wires are utilized for passage of stents, perforation of the ureter can occur, necessitating stent placement for treatment. ^{54 59} Rarely, anuria occurs from mucosal edema causing hydronephrosis, ⁵⁸ or reflex anuria without hydronephrosis. Stent placement is required for both conditions. ⁵⁹ Our group selectively utilizes PUSP for high-risk procedures and patients based on imaging and diagnosis.

Splenic Flexure Mobilization

Given the addition of operative time and injury risk for splenic flexure mobilization (SFM), surgeons should ensure necessity of mobilization and feel confident in technical ability to perform this maneuver. The role of SFM in surgery for complicated diverticulitis is colon elongation for a tension-free anastomosis. Some studies indicate length gained with SFM, ligation of IMA, and ligation of inferior mesenteric vein (IMV) at 28 to 30 cm, ^{60 61} while others indicate that only 3 to 6 cm is gained when adding SFM to high-IMA ligation and complete mobilization of the descending colon to the splenic flexure. ^{61 62} Authors suggest more length was gained by performing high ligation of the IMA, ligation of the IMV at inferior border of the pancreas, and mobilizing the entire descending colon without SFM. Further, SFM is associated with 10 to 20% increased operative time, ^{60 61 63} increased use during laparoscopy, and when diverticulitis involves the descending colon. ⁶³ Complications attributed to SFM are contradictory in the literature, with some suggesting increased rates of surgical site infection (SSI) ⁶¹ and others reporting no difference. ⁶³ Bleeding is a very uncommon complication of SFM. Our group selectively utilizes SFM in elective resection for complicated diverticulitis, and we have not seen selective use for diverticulitis increase our rate of complication.

High versus Low Ligation of the Inferior Mesenteric Artery

As lymph node harvest is unimportant (unless malignancy has not been ruled out), high ligation of the IMA is not always necessary. Other options include segmental ligation of the blood supply (low IMA and superior rectal artery [SRA] are divided with removal of V-shaped portion of mesentery) or tubular ligation (maintain IMA and SRA, instead dividing mesentery along the wall of resected colon). Some groups advocate segmental ligation, with high- or low-IMA ligation for conduit length and a tension-free anastomosis. Proponents of tubular ligation propose maintenance of the SRA blood supply that leads to fewer anastomotic complications without sacrificing a tension-free anastomosis.

In a cadaver study evaluating length of colon obtained between high- (1-cm distal to aorta) and low-IMA ligation (1-cm distal to left colic artery take-off) to create a coloanal anastomosis, the proximal sigmoid was identified where it overlay the left psoas and used to measure length gained. ⁶⁴ Division of the IMV occurred at the inferior border of the pancreas with high IMA ligation cases and at the point of left colic artery (LCA) ligation in low IMA cases. No difference in conduit length gained was seen between high- and low-IMA division alone (2.9 vs. 3.1 cm). However, IMV ligation at the pancreas border gained 8.1 cm, significantly longer than the 2.5 cm gained with IMV ligation at the LCA. Subsequent dissection of the descending colon conduit gained 8.1 cm following high-IMA ligation compared with 3.3 cm with low-IMA ligation. High ligation of the IMA with ligation of the IMV at the pancreas always reached for a coloanal anastomosis, while almost 25% of low-IMA ligation conduits could not reach. Incidentally, there was no reach advantage with IMV ligation at the inferior border of the pancreas without high-IMA ligation.

Aside from conduit length for a tension-free anastomosis, adequate blood supply may be affected by location of IMA ligation. ^{65 66} In a study of high versus low IMA ligation colectomies, a significantly higher percentage of high-ligation patients demonstrated proximal bowel necrosis postoperatively (2 vs. 0%), ⁶⁵ while another report indicated high-IMA ligation as independently associated with leak compared with low ligation (OR = 3.8). ⁶⁷ When comparing 100 tubular resections (preservation of the SRA) to 113 segmental resections for diverticulitis, a higher leak rate was noted in segmental resections (7 vs. 1%). ⁶⁸ Despite these studies, multiple other reports and a meta-analysis demonstrate no difference in leak rate between ligation or preservation of the IMA in colectomy for diverticular disease. ^{53 69 70}

Aside from affecting blood supply and leak rates, dissection of the IMA and SRA can also affect postoperative conduit function via nerve damage. In a randomized control trial of tubular resection vs low-IMA ligation in colectomy for diverticulitis, postoperative defecatory disorder (PDD), including fecal incontinence, stool fragmentation, urgency, soilage, and alternate bowel function, was evaluated at 6-month postoperative with questionnaires and anal manometry. ⁷¹ After 6 years of study, significantly lower rates of constipation, PDD, and incontinence scores were noted in the tubular resection group, producing a higher quality-of-life score despite no difference in anal manometry findings. The authors concluded preservation of the IMA and associated nerves in colectomy for diverticulitis may produce functional benefits.

Given the mixed data on segmental (high- or low-IMA ligation) versus tubular colectomy for diverticular disease, surgeons are advised to individualize patient care. We systematically review cross-sectional imaging and risk-stratifying patients based on vascular comorbidity. For patients with peripheral vascular or cardiovascular disease, preservation of the IMA and SRA are important. If tubular resection with adequate removal of disease was possible in these patients based on redundant colon length, we proceeded with resection. If length was an issue, we performed low ligation distal to the LCA, fully mobilize the conduit, and perform SFM to gain conduit length. Our last option in vasculopaths is high ligation of the IMA, which is not often necessary for diverticular disease. In patients without significant vascular disease, we attempt tubular resection. All anastomoses are end-of-rectum to side-of-colon given the improved blood supply and lower risk of leak ⁶⁷ and are evaluated with flexible sigmoidoscopy leak test.

Role of Proximal Fecal Diversion

From 2005 to 2011, only 348 (2.2%) colectomies for diverticular disease captured in NSQIP involved proximal fecal diversion. ⁷² On multivariate analysis, age >65 years, body mass index (BMI) >30 kg/m ² , tobacco abuse, steroid use, and preoperative albumin <3 mg/dL were independently associated with occurrence of proximal fecal diversion. Creation of diversion was associated with higher rates of any adverse outcome, SSI, deep-space infection, acute renal failure, and longer hospital length of stay. Similarly, studies show 17 to 45% of patients with stoma have at least one complication. ⁷³ Upwards of 3 to 25% of temporary stomas are never closed and morbidity rates of closure approach 45%. ⁷³ Given that almost 97% of diverted patients, who receive no benefit from diversion, are at higher risk for complication, and require a second operation to establish continuity; creation of an ileostomy should be carefully considered before performing. We rarely perform proximal diversion in patients with colectomy for complicated diverticulitis.

Preventing Complications: Leak, Stricture, and Recurrence

While tissue oxygen tension related to conduit blood supply was previously associated with anastomotic leak, ^{66 74} recent reports suggest that it has no effect. ⁷⁵ Despite this, a focus on evaluation of conduit bloody supply for prevention of leak and stricture is reasonable. Visual inspection of pulsatile blood flow in the marginal vessel, pulsatile flow in the arcade vessels, and bleeding at the cut or stapled edge of conduit and rectum are cost effective and highly successful. ⁷⁶ Other options include visible light spectroscopy (VLS) and near-IR spectroscopy (NIRS) using indocyanine green (ICG; (firefly, spy, etc.). ^{74 77} Unfortunately, while some suggest change in operative planning based on results from VLS and NIRS, ^{77 78} comparison studies are limited and no clear values predictive of anastomotic complication currently exist. We use palpation and visual inspection of blood supply before and after anastomosis.

Another factor contributing to anastomotic complication involves excess tension. However, knowing “how much” tension is difficult to gauge. Anecdotally, surgeons describe a “floppy” loop of colon in the pelvis as appropriate, while others suggest a conduit staple line reaching the top of the pubic symphysis as appropriate. A tensile load of only 4 g/mm ² significantly reduces submucosal blood flow to both antimesenteric and mesenteric sides of the conduit ⁷⁹ and a rapid reduction in explanted tissue integrity occurs on the mesenteric side when subjected to minimal strain of 5%. ⁸⁰ Unfortunately, mechanical determination of acute strain in the operative theater is not currently possible. Our group relies on tactile feedback, conduit staple line to pubic symphysis, and stressing of the anastomosis during leak testing with flexible sigmoidoscopy to determine if too much anastomotic tension is present.

While leak is worrisome to patient and surgeon, stricture formation can be equally troublesome. Incidence is reported as 0 to 30% with only 5% becoming symptomatic. ⁸¹ With stricture defined as lumen diameter less than 12 mm (diameter of small flexible scope) patients report symptoms of left iliac fossa pain when passing stool or flatus, abdominal distention, constipation, and thin stools. ⁸¹ In patients with this complication, 17 of 19 patients did not have prior SFM or IMA/IMV ligation at initial procedure, suggesting tension and associated ischemia as the source of stricture. ⁸¹ While anastomotic evaluation is not performed unless patients complain, studies suggest performing digital rectal examination and proctoscopy at 4 to 6 weeks postoperative preempts stricture. ⁷⁶ Distention of the anastomosis when plasticity still exists is thought to prevent stricture formation by establishing an adequate luminal diameter. In support of this, 68 consecutive laparoscopic sigmoid colectomies for diverticulitis between 1998 and 2007 were evaluated for stricture starting 3 to 5 weeks postoperatively. ⁷⁵ Almost all patients received a 31-mm stapled (93%) end-to-end (85%) anastomosis. At a mean follow-up of 226 days, 65% of patients had an anastomosis <19 mm on initial 3 to 5 weeks of proctoscopy, while 32% of patients complained of symptoms such as pain, bloating, and stool changes. Of the symptomatic patients, 45% had spontaneous resolution of narrowing and symptoms, while 11 of 12 patients with persistent stenosis and symptoms underwent successful endoscopic dilation. For long, irregular strictures, or failure of dilation following three attempts, surgical resection is indicated. If the stricture is very short (<5 mm), consider placement of stapler anvil via colotomy across the stricture followed by refiring of the transanally inserted circular stapler. ⁷⁵

Recurrence of diverticulitis following colectomy is uncommon, occurring in 7 to 16% of patients. ⁸² The average time to recurrence is approximately 45 months, yielding a 3 to 13% rate of reoperation. Failure to resect, the diseased segment down to rectum is the only independently associated risk factor for recurrence. When anastomosis is made to the rectum, rates of recurrence are less than 7 versus 13% when colon conduit is anastomosed to the distal sigmoid colon. ^{82 83} In patients who present with recurrent symptoms, a full workup is necessary to rule out other sources of pain such as irritable bowel syndrome or low anterior syndrome. Careful planning is required before reoperating for recurrent disease.

Conclusion

The diagnosis and management of complicated diverticular disease has shifted from a surgery-first approach to a more conservative drain-and-watch approach in recent years. As more studies elucidate the true risks of nonsurgical management for complicated disease, our ability to finely tune intervention for this difficult disease process will improve. For patients requiring emergent or elective surgery, a minimally invasive approach is indicated. Careful attention to blood supply, tension, and appropriate specimen resection will improve both short- and long-term patient outcomes.